Modulation of Estrogen Signaling by the Novel Interaction of Heat Shock Protein 27, a Biomarker for Atherosclerosis, and Estrogen Receptor ?
http://www.100md.com
动脉硬化血栓血管生物学 2005年第3期
From University of Ottawa Heart Institute, Ontario, Canada.
Correspondence to Dr Edward O’Brien, 40 Ruskin Street, University of Ottawa Heart Institute, Ontario, Canada K1Y4W7. E-mail eobrien@ottawaheart.ca
Abstract
Objective— We sought to discover proteins that associate with estrogen receptor beta (ER?) and modulate estrogen signaling.
Methods and Result— Using a yeast 2-hybrid screen, we identified heat shock protein 27 (HSP27) as an ER?-associated protein. HSP27 is a recently identified biomarker of atherosclerosis that is secreted at reduced levels from atherosclerotic compared with normal arteries. In vitro protein-binding assays confirmed the specific interaction of HSP27 with ER? and not ER. HSP27 expression was absent in coronary arteries with complex atherosclerotic lesions. Interestingly, HSP27 expression was also absent in 60% of coronary arteries from young males and females (27±6.5 years) with normal histology or nonobstructive fatty streaks/atheromas. Moreover, the absence of HSP27 in these normal or minimally diseased arteries coincided with the loss of ER? expression. Only 35% of arteries showed coexpression of HSP27 and ER?. Relative to controls, estradiol-mediated transcription was reduced 20% with overexpression of HSP27 and increased 44% when HSP27 protein levels were reduced with HSP27 siRNA.
Conclusions— HSP27, an ER?-associated protein, shows attenuated expression with coronary atherosclerosis and modulates estrogen signaling.
We sought to discover proteins that associate with estrogen receptor beta (ER?) and modulate estrogen signaling. We identified heat shock protein 27 (HSP27) as an ER?-associated protein and confirmed the specific interaction of HSP27 with ER? and not ER. HSP27 expression was absent in coronary arteries with complex atherosclerotic lesions. Moreover, the absence of HSP27 in these normal or minimally diseased arteries coincided with the loss of ER? expression. HSP27, an ER?-associated protein, shows attenuated expression with coronary atherosclerosis and modulates estrogen signaling.
Key Words: atherosclerosis ? hormones ? receptors ? signal transduction ? women
Introduction
There is a sex bias in the prevalence of cardiovascular morbidity and mortality that favors women until menopause; thereafter, this difference is lost.1 Although a plausible explanation for this epidemiological distinction is the presumed vasculoprotective effect of ovarian hormones, randomized primary and secondary prevention trials involving hormone replacement therapy not only are nonconfirmatory but also document ill effects.2–4 Estrogens act via at least 2 receptors that are expressed in the vessel wall (ER and ER?), although there is increasing evidence that receptor-associated proteins play a critical role in determining the biological responses to ligand-dependent and ligand-independent signaling.1,5–9
We hypothesized that coregulators of ERs may modulate estrogen signaling in vascular tissues. In this article, we report that heat shock protein 27 (HSP27), a recently reported potential biomarker of atherosclerosis, specifically interacts with estrogen receptor beta (ER?)—the receptor isoform that shows transient mRNA overexpression early after vascular injury.10–12 Using a differential proteomic screening approach, HSP27 secretion levels from human carotid plaques were found to be markedly diminished compared with normal arteries.10 Moreover, circulating blood levels of HSP27 were decreased in patients with carotid atherosclerosis relative to healthy subjects. We now demonstrate that expression of HSP27 diminishes as the stage of coronary atherosclerosis progresses and that HSP27 is capable of regulating estrogen mediated transcription in vitro.
Materials and Methods
Briefly, our studies involve 4 major components: (1) the use of a yeast 2-hybrid screen to identify proteins that associate with the unique A/B domain of ER?; (2) the in vitro confirmation of the association of HSP27 with ER? and not ER; (3) the demonstration of HSP27 expression in human coronary arteries; and (4) determination of the effect of modulating HSP27 levels on estrogen signaling. For a detailed account of the methodologies used in this article, please see http://atvb.ahajournals.org.
Results
HSP27 Interacts Specifically With the A/B Domain of ER?
Using a yeast 2-hybrid screen, we identified previously unreported protein interactions with the ER?. We selected the A/B domain of ER? as our "bait" because it shares only 30% sequence identity with ER. Dual-reporter gene activity (LEU2 and ?-galactosidase) identified 77 positive screening results from a human lung fibroblast cell line cDNA library that contained 5.3x106 clones. Restriction enzyme analysis identified duplicate clones, and the unique clones were sequenced. Because the A/B domain of ER? contains the transcriptional activation function AF1, we ensured that selection gene expression was dependent on the interaction with the candidate proteins and not simply caused by auto-activation. Discounting ferritin and ribosomal proteins, which are commonly reported false-positives (http://www.fccc.edu/research/labs/golemis/), 3 unique clones remained: B8, D6, and HSP27 (BC073768; Figure 1A and 1B). All screens were repeated in triplicate to ensure the validity of the interactions. The HSP27 clone shared 98.7% sequence identity with the HSP27 protein sequence in the open reading frame (Figure 1C).
Figure 1. Identification of HSP27 as a novel ER? associated protein. A, Yeast were transformed with an ER?–LexA DNA binding domain and cDNA clones from the library fused to an activation domain before being selected for growth on galactose (top panel) and glucose (bottom panel) in the absence of leucine (Leu–). B, ?-galactosidase activity was assessed on galactose (top) or glucose (bottom). C, Positive clone C5 was sequenced and identified as hsp27. Clone C5 shares 98.7% protein sequence identity with HSP27. D, Maltose-binding protein assay showing the interaction between the ER?–MBP fusion product with intein protein alone and HSP27-intein fusion product, using antibodies specific for intein. E, Coimmunoprecipitation performed using HeLa cells transfected with ER?–DsRed construct alone or ER?–DsRed plus hsp27–EGFP. An antibody to DsRed was used for immunoprecipitation, and an antibody to HSP27 was used for immunodetection. F, Coimmunoprecipitation of HeLa cells transfected with an ER expression vector (top panel) or an ER? expression vector (bottom panel). For coimmunoprecipitation (IP), either no antibody (no Ab) or a HSP27-specific antibody was used. Both ER-specific and ER?-specific antibodies were used for immunoblotting (IB).
To confirm our yeast 2-hybrid findings, we used 2 separate in vitro assays. First, a fusion protein pull-down assay was performed using maltose binding protein (MBP) column precipitation. Fusion proteins of ER?–MBP and HSP27–intein were expressed in bacteria, and then cell lysates were isolated. The ER?-MBP protein was immobilized in a column before either HSP27–intein or nonfused intein cell lysates were run through the column. The retained product was then isolated and analyzed by Western blot with an intein-specific probe. With the empty intein construct, minimal protein was retained in the column, indicating that intein alone interacted only weakly with the ER?–MBP product (ie, background; Figure 1D). However, the HSP27-intein fusion construct resulted in much darker band, consistent with a greater retention of protein and a strong HSP27 and ER? interaction. Second, coimmunoprecipitation of the 2 proteins was performed. HeLa cells were transfected with either an ER?-DsRed fusion construct (alone) or both ER?–DsRed and HSP27–enhanced green fluoroscent protein (EGFP) fusion protein constructs. An anti-DsRed antibody was used to immunoprecipitate proteins that bound to ER?, and isolated proteins were analyzed by Western blot analysis using an HSP27 specific antibody. In HeLa cells transfected with the ER? construct alone a weak band was observed—consistent with nonspecific binding (Figure 1E). In HeLa cells cotransfected with both the ER? and HSP27 fusion protein constructs, the 50-kDa HSP27–EGFP fusion protein was immunoprecipitated, thereby indicating a specific interaction between ER? and HSP27.
To ensure that HSP27 specifically interacted with ER? and not ER, coimmunoprecipitation of HeLa cell lysate from cells transfected with either an ER or an ER? expression vector was performed with an anti-HSP27 antibody. The precipitate was then analyzed by Western blot with both ER?-specific and ER-specific antibodies. The anti-HSP27 antibody clearly immunoprecipitated ER? but failed to demonstrate an appreciable signal above background levels for ER (Figure 1F).
HSP27 Is Coexpressed With ER? in the Coronary Arteries of Young Individuals
The expression of ER? and HSP27 was examined in coronary arteries with advanced atherosclerosis, as well as the proximal left anterior descending of 20 young human subjects who died of noncardiovascular causes (14 men and 6 women; age: 27±6.5 years). HSP27 expression was absent in those arteries with advanced atherosclerotic lesions (Figure 2A and 2B). However, in coronary arteries with benign intimal hyperplasia or nonobstructive, pathological fatty streaks and/or atheromas (Stary type I and type III lesions, respectively), we noted variable expression of both HSP27 and ER? (Figure 2C to 2F). The endothelium was present in all arteries, thereby discounting the possibility that false-negative immunolabeling results were the result of a denuded endothelium (Figure 2G). In these arteries, the following relationships were noted for the expression of HSP27 and ER?: 7 of 20 (35%) were immunopositive for both, whereas 12 of 20 (60%) were negative for both (Figure 2H). Only one artery was immunopositive for HSP27 but lacked immunodetectable ER?. No histological features were predictive of HSP27 or ER? expression in these arteries; however, on serial tissue slides the expression of HSP27 and ER? in luminal endothelial cells was clearly related (eg, R=0.94, R2=0.88; P<0.0001).
Figure 2. Expression of HSP27 in human coronary arteries. Photomicrographs of human coronary artery cross-sections: advanced atheroma with necrotic cholesterol-laden intimal core labeled for (A) -smooth muscle actin and (B) HSP27 (magnification x100; L indicates lumen; I, intima; M, media). HSP27 is absent in this atherosclerotic artery. Cross-section of a benign intimal thickening in the coronary artery of a young individual free of atherosclerosis, immunolabeled for -smooth muscle actin (C) and HSP27 (D) (magnification x400). Top left insert of (A) and (C) show Movat pentachrome-stained slides; top right insert of (C) and (D) shows immunolabeling with anti–smooth muscle -actin antibody; magnification x40. Immunodetection of ER? (E) HSP27 (F) and von Willebrand factor (vWF) (G) in normal coronary artery (magnification x1000). Hematoxylin nuclear counterstain; brown reaction product for positive immunolabeling. The chart (H) displays the frequency of immunolabeling for ER? and HSP27 in the coronary arteries of young individuals free of complex atherosclerotic disease. Apart for one exception (5%), arteries were either both immunonegative (60%) or both immunopositive (35%) for ER? and HSP27.
HSP27 Estradiol Mediated Transcriptional Activity
Finally, we examined the role of HSP27 in modulating the estrogen transcriptional activity in HeLa cells stably transfected with an estrogen response element (ERE)-driven EGFP reporter. First, we examined the effect of elevated HSP27 levels on estrogen signaling by inducing overexpression of endogenous HSP27 by treating the cells with H2O2, a form of oxidative stress (Figure 3A). 13 Induction of HSP27 overexpression (Figure 3A inset) caused a 20% decrease in 17?-estradiol transcriptional activity, as reflected by a decrease in relative fluorescence units normalized to the μg of protein per well, compared with control cells treated with 17?-estradiol but not H2O2 (P<0.05; Figure 3A). Second, we used HSP27–siRNA to reduce HSP27 protein levels in the same ERE reporter construct-transfected HeLa cells. Relative to oligofectamine or scrambled controls, siRNA dramatically reduced HSP27 protein expression (Figure 3B) and increased ERE output 44% and 59% compared with scrambled (nonsense) oligomers and oligofectamine, respectively (Figure 3C; ANOVA P<0.001, with P<0.05 for the mentioned comparisons). Taken together, these findings suggest that HSP27 is a corepressor of estrogen signaling mediated by ER?.
Figure 3. HSP27 and ERE signaling. HeLa cells stably transfected with an ERE–EGFP reporter construct were used to test the effects of elevated and decreased levels of HSP27 on estrogen signaling. All cells were treated with 100 nM 17?-estradiol. A, H2O2 treatment to induce upregulated expression of endogenous HSP27 (inset Western blot) and resulted in a 20% decrease in ERE reporter output (relative fluorescence unit [RFU]; normalized to μg of total cell protein; *P<0.05; n=12/group). B, Use of siRNA to HSP27 dramatically reduced HSP27 levels relative to scrambled (nonsense) oligomers and oligofectamine alone. GAPDH loading of each lane is shown in the lower row. C, The resultant decrease in HSP27 protein with siRNA treatment (B) produced increases in ERE output (RFU normalized to number of cells) of 44% and 59% relative to scrambled and oligofectamine; respectively (ANOVA: P<0.001, with P<0.05 for the 2 noted comparisons).
Discussion
Nuclear receptor-associated proteins are important determinants of the cellular response to ligand-dependent and ligand-independent steroid hormone signaling.1,9 Because we are interested in the vascular effects of estrogens, we sought to identify possible coregulatory proteins of ER?, the ER isoform that shows transient mRNA overexpression after vascular injury.11,12 Using a yeast 2-hybrid screen, we discovered the association of HSP27, a recently recognized biomarker of atherosclerosis, with ER?.10 By way of various in vitro protein assays, we confirmed the interaction of HSP27 with ER? but not ER. Whereas Martin-Ventura et al recently noted decreased secretion of HSP27 in the supernatant of atherosclerotic carotid plaques compared with normal human arteries, these investigators did observe HSP27 immunopositive smooth muscle cells in atherosclerotic carotid endarterectomy specimens and normal mammary arteries. In contrast, we found an absence of HSP27 expression in coronary arteries with complex atherosclerotic lesions. There are several potential explanations for this discrepancy in the HSP27 immunolabeling results: (1) different arteries were examined (carotid versus coronary); (2) perhaps the complexity of the atherosclerotic lesions differed; and (3) we used a monoclonal anti-HSP27 antibody whereas Martin-Ventura et al used a polyclonal anti-serum to HSP27. Although immunolabeling has its limitations, we were intrigued to find either an absence or presence of both HSP27 and ER? in all but 1 coronary artery from this population of young subjects. Whether the expression of these 2 proteins is linked requires further study. Interestingly, sex did not predict expression of ER? or HSP27.
Heat shock proteins are highly conserved molecular chaperones that show upregulated expression in response to a range of cellular insults (eg, heat, oxidative stress, infection, cytokines) and play an active role in the stabilization and refolding of key intracellular proteins.14 Although vascular smooth muscle and endothelial cells are known to express HSP27, the role of this protein in the vessel wall is only now beginning to be studied. Moreover, a number of studies report provocative associations between circulating heat shock protein levels or anti-heat shock protein antibodies and vascular disease.15
Even though it has been known for more than a decade that HSP27 is induced by estrogens and in some way associated with estrogen receptors in various cells (eg, breast and endometrial tumors, platelets), these studies were completed before the discovery of ER?.16,17 Our study is the first to report that HSP27 specifically associates with ER? and acts as a corepressor of estrogen signaling. Coregulatory proteins confer milieu-specific responses to steroid hormone receptors and altered levels of these factors play important roles in some diseases. For example, Gregory et al demonstrated that steroid receptor coactivators of the p160 family are expressed in the normal endometrium during the menstrual cycle but overexpressed in women with polycystic ovarian syndrome, and lead to poor reproductive function, endometrial hyperplasia, and cancer.18
Similarly, as a corepressor of estrogen signaling, HSP27 may play a role in atherogenesis. However, the critical question that needs to be resolved is whether HSP27 loss contributes directly to coronary atherogenesis or if it is merely a secondary phenomenon that follows the accumulation of an atherosclerotic plaque? Given that induction of HSP27 with herbimycin A reduces neointimal hyperplasia in rat carotid arteries subjected to balloon injury, it is attractive to postulate that the relative absence of HSP27 may in fact be an important mechanistic step in atherogenesis.19 For example, because many important vascular growth factors and cytokines have an ERE, it is conceivable that unregulated estrogen mediated transcription of these factors might occur in the absence of HSP27 and foster atherogenesis.20 Hence, knowledge of an individual’s HSP27 "status," perhaps reflected by a simple blood test, may be predictive of atherogenesis and who should receive estrogen therapy, because the development of undesirable side effects (eg, venous thrombosis, malignancy) could be caused by loss of HSP27 regulation of estrogen-mediated transcription. With studies of HSP27 ongoing in various patient populations, the usefulness of this biomarker in clinical management will soon be clarified.
Acknowledgments
This work was supported by a grant-in-aid (#T5073; to E.O.B.) from the Heart and Stroke Foundation of Ontario (HSFO). Also, HSFO program grant #5275 funded some of the laboratory equipment used in this study. E.O.B. is a CIHR-University Industry Investigator. The authors are indebted to Erik Holm for his assistance with the preparation of this manuscript.
Received September 15, 2004; accepted December 28, 2004.
References
Turgeon JL, McDonnell DP, Martin KA, Wise PM. Hormone Therapy: Physiological Complexity Belies Therapeutic Simplicity. Science. 2004; 304: 1269–1273.
Hulley S, Grady D, Bush T, Furberg C, Herrington D, Riggs B, Vittinghoff E, for the Heart and Estrogen/progestin Replacement Study Research Group. Randomized Trial of Estrogen Plus Progestin for Secondary Prevention of Coronary Heart Disease in Postmenopausal Women. JAMA. 1998; 280: 605–613.
Writing Group for the Women’s Health Initiative Investigators. Risks and Benefits of Estrogen Plus Progestin in Healthy Postmenopausal Women: Principal Results From the Women’s Health Initiative Randomized Controlled Trial. JAMA. 2002; 288: 321–333.
Waters DD, Gordon D, Rossouw JE, Cannon RO, III, Collins P, Herrington DM, Hsia J, Langer R, Mosca L, Ouyang P, Sopko G, Stefanick ML, Endorsed by the Am College of Cardiology Foundation. Women’s Ischemic Syndrome Evaluation: Current Status and Future Research Directions. Report of the National Heart, Lung and Blood Institute Workshop: Lessons From Hormone Replacement Trials. Circulation. 2004; 109: 53–55.
Greene GL, Sobel NB, King WJ, Jensen EV. Immunochemical studies of estrogen receptors. J Steroid Biochem Mol Biol. 1984; 20: 51–56.
Kuiper GGJM, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson J-A. Cloning of a novel estrogen receptor expressed in rat prostrate and ovary. Proc Natl Acad Sci U S A. 1996; 93: 5925–5930.
Mosselman S, Polman J, Dijkema R. ERbeta: Identification and characterization of a novel human estrogen receptor. FEBS. 1996; 392: 49–53.
Halachmi S, Marden E, Martin G, MacKay H, Abbondanza C, Brown M. Estrogen receptor - associated proteins: possible mediators of hormone-induced transcription. Science. 1994; 264: 1455–1458.
Smith CL, O’Malley BW. Coregulator Function: A Key to Understanding Tissue Specificity of Selective Receptor Modulators. Endocr Rev. 2004; 25: 45–71.
Martin-Ventura JL, Duran MC, Blanco-Colio LM, Meilhac O, Leclercq A, Michel JB, Jensen ON, Hernandez-Merida S, Tunon J, Vivanco F, Egido J. Identification by a differential proteomic approach of heat shock protein 27 as a potential marker of atherosclerosis. Circulation. 2004; 110: 2216–2219.
Lindner V, Kim SK, Karas RH, Kuiper GJM, Gustafsson J-A, Mendelsohn ME. Increased expression of estrogen receptor-beta mRNA in male blood vessels after vascular injury. Circ Res. 1998; 83: 224–229.
Makela S, Savolainen H, Aavik E, Myllarniemi M, Strauss L, Taskinen E, Gustafsson JA, Hayry P. Differentiation between vasculoprotective and uterotrophic effects of ligands with different binding affinities to estrogen receptors alpha and beta. Proc Natl Acad Sci U S A. 1999; 96: 7077–7082.
Mehlen P, Briolay J, Smith L, az-latoud C, Fabre N, Pauli D, Arrigo AP. Analysis of the resistance to heat and hydrogen peroxide stresses in COS cells transiently expressing wild type or deletion mutants of the Drosophila 27-kDa heat-shock protein. Eur J Biochem. 1993; 215: 277–284.
Benjamin IJ, McMillan DR. Stress (heat shock) proteins: molecular chaperones in cardiovascular biology and disease. Circ Res. 1998; 83: 117–132.
Pockley AG. Heat shock proteins, inflammation, and cardiovascular disease. Circulation. 2002; 105: 1012–1017.
Ciocca DR, Oesterreich S, Chamness GC, McGuire WL, Fuqua SA. Biological and clinical implications of heat shock protein 27,000 (Hsp27): a review. J Natl Cancer Inst. 1993; 85: 1558–1570.
Mendelsohn ME, Zhu Y, O’Neill S. The 29-kDa proteins phosphorylated in thrombin-activated human platelets are forms of the estrogen receptor-related 27-kDa heat shock protein. Proc Natl Acad Sci. 1991; 88: 11212–11216.
Gregory CW, Wilson EM, Apparao KBC, Lininger RA, Meyer WR, Kowalik A, Fritz MA, Lessey BA. Steroid Receptor Coactivator Expression throughout the Menstrual Cycle in Normal and Abnormal Endometrium. J Clin Endocrinol Metab. 2002; 87: 2960–2966.
Connolly EM, Kelly CJ, Chen G, O’grady T, Kay E, Leahy A, Bouchier-Hayes DJ. Pharmacological induction of HSP27 attenuates intimal hyperplasia in vivo. Eur J Vasc Endovasc Surg. 2003; 25: 40–47.
Mendelsohn ME, Karas RH. The protective effects of estrogen on the cardiovasclar system. N Eng J Med. 1999; 340: 1801–1811.(Harvey Miller; Stephanie )
Correspondence to Dr Edward O’Brien, 40 Ruskin Street, University of Ottawa Heart Institute, Ontario, Canada K1Y4W7. E-mail eobrien@ottawaheart.ca
Abstract
Objective— We sought to discover proteins that associate with estrogen receptor beta (ER?) and modulate estrogen signaling.
Methods and Result— Using a yeast 2-hybrid screen, we identified heat shock protein 27 (HSP27) as an ER?-associated protein. HSP27 is a recently identified biomarker of atherosclerosis that is secreted at reduced levels from atherosclerotic compared with normal arteries. In vitro protein-binding assays confirmed the specific interaction of HSP27 with ER? and not ER. HSP27 expression was absent in coronary arteries with complex atherosclerotic lesions. Interestingly, HSP27 expression was also absent in 60% of coronary arteries from young males and females (27±6.5 years) with normal histology or nonobstructive fatty streaks/atheromas. Moreover, the absence of HSP27 in these normal or minimally diseased arteries coincided with the loss of ER? expression. Only 35% of arteries showed coexpression of HSP27 and ER?. Relative to controls, estradiol-mediated transcription was reduced 20% with overexpression of HSP27 and increased 44% when HSP27 protein levels were reduced with HSP27 siRNA.
Conclusions— HSP27, an ER?-associated protein, shows attenuated expression with coronary atherosclerosis and modulates estrogen signaling.
We sought to discover proteins that associate with estrogen receptor beta (ER?) and modulate estrogen signaling. We identified heat shock protein 27 (HSP27) as an ER?-associated protein and confirmed the specific interaction of HSP27 with ER? and not ER. HSP27 expression was absent in coronary arteries with complex atherosclerotic lesions. Moreover, the absence of HSP27 in these normal or minimally diseased arteries coincided with the loss of ER? expression. HSP27, an ER?-associated protein, shows attenuated expression with coronary atherosclerosis and modulates estrogen signaling.
Key Words: atherosclerosis ? hormones ? receptors ? signal transduction ? women
Introduction
There is a sex bias in the prevalence of cardiovascular morbidity and mortality that favors women until menopause; thereafter, this difference is lost.1 Although a plausible explanation for this epidemiological distinction is the presumed vasculoprotective effect of ovarian hormones, randomized primary and secondary prevention trials involving hormone replacement therapy not only are nonconfirmatory but also document ill effects.2–4 Estrogens act via at least 2 receptors that are expressed in the vessel wall (ER and ER?), although there is increasing evidence that receptor-associated proteins play a critical role in determining the biological responses to ligand-dependent and ligand-independent signaling.1,5–9
We hypothesized that coregulators of ERs may modulate estrogen signaling in vascular tissues. In this article, we report that heat shock protein 27 (HSP27), a recently reported potential biomarker of atherosclerosis, specifically interacts with estrogen receptor beta (ER?)—the receptor isoform that shows transient mRNA overexpression early after vascular injury.10–12 Using a differential proteomic screening approach, HSP27 secretion levels from human carotid plaques were found to be markedly diminished compared with normal arteries.10 Moreover, circulating blood levels of HSP27 were decreased in patients with carotid atherosclerosis relative to healthy subjects. We now demonstrate that expression of HSP27 diminishes as the stage of coronary atherosclerosis progresses and that HSP27 is capable of regulating estrogen mediated transcription in vitro.
Materials and Methods
Briefly, our studies involve 4 major components: (1) the use of a yeast 2-hybrid screen to identify proteins that associate with the unique A/B domain of ER?; (2) the in vitro confirmation of the association of HSP27 with ER? and not ER; (3) the demonstration of HSP27 expression in human coronary arteries; and (4) determination of the effect of modulating HSP27 levels on estrogen signaling. For a detailed account of the methodologies used in this article, please see http://atvb.ahajournals.org.
Results
HSP27 Interacts Specifically With the A/B Domain of ER?
Using a yeast 2-hybrid screen, we identified previously unreported protein interactions with the ER?. We selected the A/B domain of ER? as our "bait" because it shares only 30% sequence identity with ER. Dual-reporter gene activity (LEU2 and ?-galactosidase) identified 77 positive screening results from a human lung fibroblast cell line cDNA library that contained 5.3x106 clones. Restriction enzyme analysis identified duplicate clones, and the unique clones were sequenced. Because the A/B domain of ER? contains the transcriptional activation function AF1, we ensured that selection gene expression was dependent on the interaction with the candidate proteins and not simply caused by auto-activation. Discounting ferritin and ribosomal proteins, which are commonly reported false-positives (http://www.fccc.edu/research/labs/golemis/), 3 unique clones remained: B8, D6, and HSP27 (BC073768; Figure 1A and 1B). All screens were repeated in triplicate to ensure the validity of the interactions. The HSP27 clone shared 98.7% sequence identity with the HSP27 protein sequence in the open reading frame (Figure 1C).
Figure 1. Identification of HSP27 as a novel ER? associated protein. A, Yeast were transformed with an ER?–LexA DNA binding domain and cDNA clones from the library fused to an activation domain before being selected for growth on galactose (top panel) and glucose (bottom panel) in the absence of leucine (Leu–). B, ?-galactosidase activity was assessed on galactose (top) or glucose (bottom). C, Positive clone C5 was sequenced and identified as hsp27. Clone C5 shares 98.7% protein sequence identity with HSP27. D, Maltose-binding protein assay showing the interaction between the ER?–MBP fusion product with intein protein alone and HSP27-intein fusion product, using antibodies specific for intein. E, Coimmunoprecipitation performed using HeLa cells transfected with ER?–DsRed construct alone or ER?–DsRed plus hsp27–EGFP. An antibody to DsRed was used for immunoprecipitation, and an antibody to HSP27 was used for immunodetection. F, Coimmunoprecipitation of HeLa cells transfected with an ER expression vector (top panel) or an ER? expression vector (bottom panel). For coimmunoprecipitation (IP), either no antibody (no Ab) or a HSP27-specific antibody was used. Both ER-specific and ER?-specific antibodies were used for immunoblotting (IB).
To confirm our yeast 2-hybrid findings, we used 2 separate in vitro assays. First, a fusion protein pull-down assay was performed using maltose binding protein (MBP) column precipitation. Fusion proteins of ER?–MBP and HSP27–intein were expressed in bacteria, and then cell lysates were isolated. The ER?-MBP protein was immobilized in a column before either HSP27–intein or nonfused intein cell lysates were run through the column. The retained product was then isolated and analyzed by Western blot with an intein-specific probe. With the empty intein construct, minimal protein was retained in the column, indicating that intein alone interacted only weakly with the ER?–MBP product (ie, background; Figure 1D). However, the HSP27-intein fusion construct resulted in much darker band, consistent with a greater retention of protein and a strong HSP27 and ER? interaction. Second, coimmunoprecipitation of the 2 proteins was performed. HeLa cells were transfected with either an ER?-DsRed fusion construct (alone) or both ER?–DsRed and HSP27–enhanced green fluoroscent protein (EGFP) fusion protein constructs. An anti-DsRed antibody was used to immunoprecipitate proteins that bound to ER?, and isolated proteins were analyzed by Western blot analysis using an HSP27 specific antibody. In HeLa cells transfected with the ER? construct alone a weak band was observed—consistent with nonspecific binding (Figure 1E). In HeLa cells cotransfected with both the ER? and HSP27 fusion protein constructs, the 50-kDa HSP27–EGFP fusion protein was immunoprecipitated, thereby indicating a specific interaction between ER? and HSP27.
To ensure that HSP27 specifically interacted with ER? and not ER, coimmunoprecipitation of HeLa cell lysate from cells transfected with either an ER or an ER? expression vector was performed with an anti-HSP27 antibody. The precipitate was then analyzed by Western blot with both ER?-specific and ER-specific antibodies. The anti-HSP27 antibody clearly immunoprecipitated ER? but failed to demonstrate an appreciable signal above background levels for ER (Figure 1F).
HSP27 Is Coexpressed With ER? in the Coronary Arteries of Young Individuals
The expression of ER? and HSP27 was examined in coronary arteries with advanced atherosclerosis, as well as the proximal left anterior descending of 20 young human subjects who died of noncardiovascular causes (14 men and 6 women; age: 27±6.5 years). HSP27 expression was absent in those arteries with advanced atherosclerotic lesions (Figure 2A and 2B). However, in coronary arteries with benign intimal hyperplasia or nonobstructive, pathological fatty streaks and/or atheromas (Stary type I and type III lesions, respectively), we noted variable expression of both HSP27 and ER? (Figure 2C to 2F). The endothelium was present in all arteries, thereby discounting the possibility that false-negative immunolabeling results were the result of a denuded endothelium (Figure 2G). In these arteries, the following relationships were noted for the expression of HSP27 and ER?: 7 of 20 (35%) were immunopositive for both, whereas 12 of 20 (60%) were negative for both (Figure 2H). Only one artery was immunopositive for HSP27 but lacked immunodetectable ER?. No histological features were predictive of HSP27 or ER? expression in these arteries; however, on serial tissue slides the expression of HSP27 and ER? in luminal endothelial cells was clearly related (eg, R=0.94, R2=0.88; P<0.0001).
Figure 2. Expression of HSP27 in human coronary arteries. Photomicrographs of human coronary artery cross-sections: advanced atheroma with necrotic cholesterol-laden intimal core labeled for (A) -smooth muscle actin and (B) HSP27 (magnification x100; L indicates lumen; I, intima; M, media). HSP27 is absent in this atherosclerotic artery. Cross-section of a benign intimal thickening in the coronary artery of a young individual free of atherosclerosis, immunolabeled for -smooth muscle actin (C) and HSP27 (D) (magnification x400). Top left insert of (A) and (C) show Movat pentachrome-stained slides; top right insert of (C) and (D) shows immunolabeling with anti–smooth muscle -actin antibody; magnification x40. Immunodetection of ER? (E) HSP27 (F) and von Willebrand factor (vWF) (G) in normal coronary artery (magnification x1000). Hematoxylin nuclear counterstain; brown reaction product for positive immunolabeling. The chart (H) displays the frequency of immunolabeling for ER? and HSP27 in the coronary arteries of young individuals free of complex atherosclerotic disease. Apart for one exception (5%), arteries were either both immunonegative (60%) or both immunopositive (35%) for ER? and HSP27.
HSP27 Estradiol Mediated Transcriptional Activity
Finally, we examined the role of HSP27 in modulating the estrogen transcriptional activity in HeLa cells stably transfected with an estrogen response element (ERE)-driven EGFP reporter. First, we examined the effect of elevated HSP27 levels on estrogen signaling by inducing overexpression of endogenous HSP27 by treating the cells with H2O2, a form of oxidative stress (Figure 3A). 13 Induction of HSP27 overexpression (Figure 3A inset) caused a 20% decrease in 17?-estradiol transcriptional activity, as reflected by a decrease in relative fluorescence units normalized to the μg of protein per well, compared with control cells treated with 17?-estradiol but not H2O2 (P<0.05; Figure 3A). Second, we used HSP27–siRNA to reduce HSP27 protein levels in the same ERE reporter construct-transfected HeLa cells. Relative to oligofectamine or scrambled controls, siRNA dramatically reduced HSP27 protein expression (Figure 3B) and increased ERE output 44% and 59% compared with scrambled (nonsense) oligomers and oligofectamine, respectively (Figure 3C; ANOVA P<0.001, with P<0.05 for the mentioned comparisons). Taken together, these findings suggest that HSP27 is a corepressor of estrogen signaling mediated by ER?.
Figure 3. HSP27 and ERE signaling. HeLa cells stably transfected with an ERE–EGFP reporter construct were used to test the effects of elevated and decreased levels of HSP27 on estrogen signaling. All cells were treated with 100 nM 17?-estradiol. A, H2O2 treatment to induce upregulated expression of endogenous HSP27 (inset Western blot) and resulted in a 20% decrease in ERE reporter output (relative fluorescence unit [RFU]; normalized to μg of total cell protein; *P<0.05; n=12/group). B, Use of siRNA to HSP27 dramatically reduced HSP27 levels relative to scrambled (nonsense) oligomers and oligofectamine alone. GAPDH loading of each lane is shown in the lower row. C, The resultant decrease in HSP27 protein with siRNA treatment (B) produced increases in ERE output (RFU normalized to number of cells) of 44% and 59% relative to scrambled and oligofectamine; respectively (ANOVA: P<0.001, with P<0.05 for the 2 noted comparisons).
Discussion
Nuclear receptor-associated proteins are important determinants of the cellular response to ligand-dependent and ligand-independent steroid hormone signaling.1,9 Because we are interested in the vascular effects of estrogens, we sought to identify possible coregulatory proteins of ER?, the ER isoform that shows transient mRNA overexpression after vascular injury.11,12 Using a yeast 2-hybrid screen, we discovered the association of HSP27, a recently recognized biomarker of atherosclerosis, with ER?.10 By way of various in vitro protein assays, we confirmed the interaction of HSP27 with ER? but not ER. Whereas Martin-Ventura et al recently noted decreased secretion of HSP27 in the supernatant of atherosclerotic carotid plaques compared with normal human arteries, these investigators did observe HSP27 immunopositive smooth muscle cells in atherosclerotic carotid endarterectomy specimens and normal mammary arteries. In contrast, we found an absence of HSP27 expression in coronary arteries with complex atherosclerotic lesions. There are several potential explanations for this discrepancy in the HSP27 immunolabeling results: (1) different arteries were examined (carotid versus coronary); (2) perhaps the complexity of the atherosclerotic lesions differed; and (3) we used a monoclonal anti-HSP27 antibody whereas Martin-Ventura et al used a polyclonal anti-serum to HSP27. Although immunolabeling has its limitations, we were intrigued to find either an absence or presence of both HSP27 and ER? in all but 1 coronary artery from this population of young subjects. Whether the expression of these 2 proteins is linked requires further study. Interestingly, sex did not predict expression of ER? or HSP27.
Heat shock proteins are highly conserved molecular chaperones that show upregulated expression in response to a range of cellular insults (eg, heat, oxidative stress, infection, cytokines) and play an active role in the stabilization and refolding of key intracellular proteins.14 Although vascular smooth muscle and endothelial cells are known to express HSP27, the role of this protein in the vessel wall is only now beginning to be studied. Moreover, a number of studies report provocative associations between circulating heat shock protein levels or anti-heat shock protein antibodies and vascular disease.15
Even though it has been known for more than a decade that HSP27 is induced by estrogens and in some way associated with estrogen receptors in various cells (eg, breast and endometrial tumors, platelets), these studies were completed before the discovery of ER?.16,17 Our study is the first to report that HSP27 specifically associates with ER? and acts as a corepressor of estrogen signaling. Coregulatory proteins confer milieu-specific responses to steroid hormone receptors and altered levels of these factors play important roles in some diseases. For example, Gregory et al demonstrated that steroid receptor coactivators of the p160 family are expressed in the normal endometrium during the menstrual cycle but overexpressed in women with polycystic ovarian syndrome, and lead to poor reproductive function, endometrial hyperplasia, and cancer.18
Similarly, as a corepressor of estrogen signaling, HSP27 may play a role in atherogenesis. However, the critical question that needs to be resolved is whether HSP27 loss contributes directly to coronary atherogenesis or if it is merely a secondary phenomenon that follows the accumulation of an atherosclerotic plaque? Given that induction of HSP27 with herbimycin A reduces neointimal hyperplasia in rat carotid arteries subjected to balloon injury, it is attractive to postulate that the relative absence of HSP27 may in fact be an important mechanistic step in atherogenesis.19 For example, because many important vascular growth factors and cytokines have an ERE, it is conceivable that unregulated estrogen mediated transcription of these factors might occur in the absence of HSP27 and foster atherogenesis.20 Hence, knowledge of an individual’s HSP27 "status," perhaps reflected by a simple blood test, may be predictive of atherogenesis and who should receive estrogen therapy, because the development of undesirable side effects (eg, venous thrombosis, malignancy) could be caused by loss of HSP27 regulation of estrogen-mediated transcription. With studies of HSP27 ongoing in various patient populations, the usefulness of this biomarker in clinical management will soon be clarified.
Acknowledgments
This work was supported by a grant-in-aid (#T5073; to E.O.B.) from the Heart and Stroke Foundation of Ontario (HSFO). Also, HSFO program grant #5275 funded some of the laboratory equipment used in this study. E.O.B. is a CIHR-University Industry Investigator. The authors are indebted to Erik Holm for his assistance with the preparation of this manuscript.
Received September 15, 2004; accepted December 28, 2004.
References
Turgeon JL, McDonnell DP, Martin KA, Wise PM. Hormone Therapy: Physiological Complexity Belies Therapeutic Simplicity. Science. 2004; 304: 1269–1273.
Hulley S, Grady D, Bush T, Furberg C, Herrington D, Riggs B, Vittinghoff E, for the Heart and Estrogen/progestin Replacement Study Research Group. Randomized Trial of Estrogen Plus Progestin for Secondary Prevention of Coronary Heart Disease in Postmenopausal Women. JAMA. 1998; 280: 605–613.
Writing Group for the Women’s Health Initiative Investigators. Risks and Benefits of Estrogen Plus Progestin in Healthy Postmenopausal Women: Principal Results From the Women’s Health Initiative Randomized Controlled Trial. JAMA. 2002; 288: 321–333.
Waters DD, Gordon D, Rossouw JE, Cannon RO, III, Collins P, Herrington DM, Hsia J, Langer R, Mosca L, Ouyang P, Sopko G, Stefanick ML, Endorsed by the Am College of Cardiology Foundation. Women’s Ischemic Syndrome Evaluation: Current Status and Future Research Directions. Report of the National Heart, Lung and Blood Institute Workshop: Lessons From Hormone Replacement Trials. Circulation. 2004; 109: 53–55.
Greene GL, Sobel NB, King WJ, Jensen EV. Immunochemical studies of estrogen receptors. J Steroid Biochem Mol Biol. 1984; 20: 51–56.
Kuiper GGJM, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson J-A. Cloning of a novel estrogen receptor expressed in rat prostrate and ovary. Proc Natl Acad Sci U S A. 1996; 93: 5925–5930.
Mosselman S, Polman J, Dijkema R. ERbeta: Identification and characterization of a novel human estrogen receptor. FEBS. 1996; 392: 49–53.
Halachmi S, Marden E, Martin G, MacKay H, Abbondanza C, Brown M. Estrogen receptor - associated proteins: possible mediators of hormone-induced transcription. Science. 1994; 264: 1455–1458.
Smith CL, O’Malley BW. Coregulator Function: A Key to Understanding Tissue Specificity of Selective Receptor Modulators. Endocr Rev. 2004; 25: 45–71.
Martin-Ventura JL, Duran MC, Blanco-Colio LM, Meilhac O, Leclercq A, Michel JB, Jensen ON, Hernandez-Merida S, Tunon J, Vivanco F, Egido J. Identification by a differential proteomic approach of heat shock protein 27 as a potential marker of atherosclerosis. Circulation. 2004; 110: 2216–2219.
Lindner V, Kim SK, Karas RH, Kuiper GJM, Gustafsson J-A, Mendelsohn ME. Increased expression of estrogen receptor-beta mRNA in male blood vessels after vascular injury. Circ Res. 1998; 83: 224–229.
Makela S, Savolainen H, Aavik E, Myllarniemi M, Strauss L, Taskinen E, Gustafsson JA, Hayry P. Differentiation between vasculoprotective and uterotrophic effects of ligands with different binding affinities to estrogen receptors alpha and beta. Proc Natl Acad Sci U S A. 1999; 96: 7077–7082.
Mehlen P, Briolay J, Smith L, az-latoud C, Fabre N, Pauli D, Arrigo AP. Analysis of the resistance to heat and hydrogen peroxide stresses in COS cells transiently expressing wild type or deletion mutants of the Drosophila 27-kDa heat-shock protein. Eur J Biochem. 1993; 215: 277–284.
Benjamin IJ, McMillan DR. Stress (heat shock) proteins: molecular chaperones in cardiovascular biology and disease. Circ Res. 1998; 83: 117–132.
Pockley AG. Heat shock proteins, inflammation, and cardiovascular disease. Circulation. 2002; 105: 1012–1017.
Ciocca DR, Oesterreich S, Chamness GC, McGuire WL, Fuqua SA. Biological and clinical implications of heat shock protein 27,000 (Hsp27): a review. J Natl Cancer Inst. 1993; 85: 1558–1570.
Mendelsohn ME, Zhu Y, O’Neill S. The 29-kDa proteins phosphorylated in thrombin-activated human platelets are forms of the estrogen receptor-related 27-kDa heat shock protein. Proc Natl Acad Sci. 1991; 88: 11212–11216.
Gregory CW, Wilson EM, Apparao KBC, Lininger RA, Meyer WR, Kowalik A, Fritz MA, Lessey BA. Steroid Receptor Coactivator Expression throughout the Menstrual Cycle in Normal and Abnormal Endometrium. J Clin Endocrinol Metab. 2002; 87: 2960–2966.
Connolly EM, Kelly CJ, Chen G, O’grady T, Kay E, Leahy A, Bouchier-Hayes DJ. Pharmacological induction of HSP27 attenuates intimal hyperplasia in vivo. Eur J Vasc Endovasc Surg. 2003; 25: 40–47.
Mendelsohn ME, Karas RH. The protective effects of estrogen on the cardiovasclar system. N Eng J Med. 1999; 340: 1801–1811.(Harvey Miller; Stephanie )